Introduction: The benefit of glycemic control was demonstrated in previous studies for the treatment of inpatients with acute infections, regardless of the diagnosis of diabetes mellitus (DM). DM is considered an important risk factor for poor outcomes in hospitalized patients. Hypothesis: This study aimed to determine the association between glycemic variability and mortality in COVID-19 hospitalized patients. Methods: This prospective study analyzed the data from patients with confirmed COVID-19, admitted to a private reference hospital, from February 2020 until July 2021 aged > 18 years. They were stratified according to the glycemic variability measured by coefficient of variation (CV) and standard deviation (SD). A logistic regression model was used to determine the association of poor glycemic variability and mortality in diabetic and non-diabetic hospitalized patients with COVID-19. Results: Data from 628 patients with confirmed COVID-19 and point-of-care (POC) glucose testing was used to measure SD and CV. The cut-off points for SD ≥ 44.7 and CV ≥ 27.5% was defined using the point of a ROC curve that predicted higher risk of death. Patients were stratified in groups for each used metric: SD < 44.7(n=357) and ≥ 44.7(n=271); CV <27.5% (n=318) and ≥ 27.5% (n=310). When controlling for multiple variables (age, sex, comorbidities and in-hospital corticoid use) there was a significant association between high VG and mortality (SD: OR 3.01 and CV: OR 2.45) (all p values < 0.05). When they were stratified as with previous DM condition and without-DM, after adjusted variables, the risk of mortality was higher in patients with uncontrolled VG, regardless the presence of DM (p<0.001). Conclusion: In conclusion, our data suggest that a grate glycemic variability is associated with higher mortality in inpatients with COVID-19 regardless of diabetes status. Efforts should be made to improve the glycemic control of these patients.